Following a near-9% drop on Friday in response to its Q2 report and earnings call remarks, Intel's (INTC) shares are pricing in a fair amount of bad news.
However, thanks to manufacturing issues, tougher competition and potential slowdowns in enterprise PC and server demand, it's quite possible that more bad news will arrive in the short-to-intermediate term, and prevent the chip giant from putting in a bottom for now.
After the bell on Thursday, Intel reported Q2 revenue of $16.96 billion (up 15% annually) and non-GAAP EPS of $1.04 (up 44%). Those numbers were slightly better than analyst estimates, as well as the guidance Intel provided on June 21st, when it disclosed CEO Brian Krzanich had resigned due to a code of conduct violation.
Intel also guided for Q3 revenue of $17.6 billion to $18.1 billion and EPS of $1.10 to $1.20, which is favorable to consensus estimates of $17.64 billion and $1.08. And it's now guiding for full-year revenue of $68.5 billion to $70.5 billion and EPS of $4.15 (plus or minus 5%), favorable to consensus estimates of $68.35 billion and $4.01.
So why did Intel's shares nosedive post-earnings? There are a few different reasons.
One of the big factors is that while Intel's total Q2 revenue beat estimates with the help of stronger-than-expected PC CPU sales, its Data Center Group (DCG), which supplies server CPUs and other data center products, missed estimates. DCG revenue of $5.55 billion, though up 27% annually with the help of strong cloud and telco demand, fell short of a $5.64 billion consensus.
Moreover, on the call, Intel guided for DCG revenue, which rose 25% over the first six months of the year, to be up only 20% for the whole of 2018. That implies a meaningful second-half growth slowdown.
Slower enterprise server sales growth could affect DCG over the next two quarters; though DCG's enterprise and government sales were up 10% in Q2, Intel expects them to decline over the long-term as workloads continue moving to public clouds.
Competition from AMD's (AMD) Epyc server CPU line could be playing some role here. On Wednesday, AMD disclosed that its Epyc revenue rose over 50% sequentially off a small base, and reiterated a goal of achieving a mid-single digit server CPU share by year's end. It's also worth noting that IBM (IBM) , whose Power and System z (mainframe) server lines rely on proprietary CPUs, saw better-than-expected hardware sales in Q2 with the help of a strong mainframe upgrade cycle.
Another factor: Intel's guidance indicates its gross margin (GM) will be pressured during the next two quarters. Whereas non-GAAP GM was flat annually in Q2 at 63%, Intel expects it to be down a percentage point in Q3 from a year-ago level of 63.9%. And during the Q&A part of the call, Bernstein analyst Stacy Rasgon estimated Intel's full-year guidance implies a Q4 GM of about 59%, which would be well below a Q4 2017 level of 64.8%.
Rasgon, who downgraded Intel to Underperform shortly after Krzanich's ouster, also thinks elevated capital spending could -- via the depreciation expenses eventually recorded for capex -- impact GMs in 2019. Intel used its Q2 report to hike its 2018 capex budget for the second time in three months; it now expects capex of $14.5 billion to $15.5 billion, well above a 2017 level of $11.8 billion.
CFO and interim CEO Bob Swan partly blamed the expected second-half margin decline on a revenue mix shift towards modems and memory chips -- Intel is expected to be the sole modem supplier for Apple's (AAPL) 2018 iPhones, after having shared modem supply duties with Qualcomm (QCOM) the last two years. He also mentioned that Intel's efforts to ramp its much-delayed 10-nanometer (10nm) manufacturing process will ding margins, and -- notably -- that Intel isn't counting on seeing the same kind of average selling price (ASP) strength it witnessed during the first half of 2018.
ASPs provided a major revenue boost to both DCG and Intel's Client Computing Group (CCG - it supplies PC and mobile chips) in Q2, and also propped up margins. DCG's ASP rose 11%, while CCG reported a 13% ASP increase for desktops and a 2% increase for notebooks.
Growing PC and server competition from AMD could impact ASPs a bit going forward. So could diminishing benefits from recent trends that have boosted ASPs, such as a PC mix shift towards business systems and last summer's launch of the well-received Xeon Scalable server CPU family.
The third factor behind the selloff: Three months after having pushed back its target for the start of 10nm mass-production (once expected by 2016) from the second half of 2018 to some point in 2019, Intel used its call to say that it only expects PCs containing 10nm chips to be "on shelves for the 2019 holiday season," and that 10nm data center products will launch afterwards.
As I mentioned after Intel's Q1 report, its 10nm delays create a golden opportunity for AMD and certain other rivals relying on Asian chip contract manufacturers (foundries) to gain ground, by launching chips based on 7nm processes that are competitive with Intel's 10nm process in late 2018 and early 2019. AMD has begun sampling both a 7nm Epyc CPU and a 7nm server GPU that are made by Taiwan Semiconductor (TSM) . Apple, Qualcomm and others are also believed to be using TSMC's 7nm process for products that will be available later this year or in the first half of 2019.
Now that opportunity to take share in 2019 looks even stronger. And while Intel might do a better job of rolling out its 7nm process (a date for starting mass-production hasn't yet been announced) than it has done with its 10nm process, the fact that TSMC plans to start mass-production in early 2020 for a 5nm process that's expected to be competitive with Intel's 7nm process means that the company doesn't have much of a margin of error here.
The good news for Intel investors: With shares now trading for just 11 times a 2019 EPS consensus of $4.22, markets have already priced in a lot of concerns about competition, margins and growth. And while recent strength in enterprise PC and server demand might not last, cloud and telco server demand should remain a long-term growth driver. So should, to varying degrees, Intel's memory, automotive and IoT opportunities.
That said, Intel's 2019 consensus estimates still imply a modest amount of revenue and EPS growth relative to 2018. And given the competitive environment the company will be facing thanks to its manufacturing issues and other factors, it's not a given that the company will hit those estimates.
Should Intel's 2019 estimates need to be taken down thanks to competitive and/or margin pressures, that could easily lead its shares, which are still up 36% over the last 12 months, to drop farther before its growth opportunities, and perhaps also its 7nm efforts, help it put in a bottom.